Coronavirus US - All-Cause Death
davegoblue
2022-06-23
Background
This module extends code contained in Coronavirus_Statistics_v005.Rmd to include sourcing of updated functions and parameters. This file includes the latest code for analyzing all-cause death data from CDC Weekly Deaths by Jurisdiction. CDC maintains data on deaths by week, age cohort, and state in the US. Downloaded data are unique by state, epidemiological week, year, age, and type (actual vs. predicted/projected).
These data are known to have a lag between death and reporting, and the CDC back-correct to report deaths at the time the death occurred even if the death is reported in following weeks. This means totals for recent weeks tend to run low (lag), and the CDC run a projection of the expected total number of deaths given the historical lag times. Per other analysts on the internet, there is currently significant supra-lag, with lag times much longer than historical averages causing CDC projected deaths for recent weeks to be low.
The code leverages tidyverse and sourced functions throughout:
# All functions assume that tidyverse and its components are loaded and available
library(tidyverse)## -- Attaching packages --------------------------------------- tidyverse 1.3.1 --## v ggplot2 3.3.3 v purrr 0.3.4
## v tibble 3.1.1 v dplyr 1.0.6
## v tidyr 1.1.3 v stringr 1.4.0
## v readr 1.4.0 v forcats 0.5.1## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()# If the same function is in both files, use the version from the more specific source
source("./Generic_Added_Utility_Functions_202105_v001.R")
source("./Coronavirus_CDC_Excess_Functions_v001.R")The basic process includes three data update steps:
# STEP 0: Appropriate parameters for 2022 data
cdcExcessParams <- list(remapVars=c('Jurisdiction'='fullState',
'Week Ending Date'='weekEnding',
'State Abbreviation'='state',
'Age Group'='age',
'Number of Deaths'='deaths',
'Time Period'='period',
'Year'='year',
'Week'='week'
),
colTypes="ccciicdcccc",
ageLevels=c("Under 25 years",
"25-44 years",
"45-64 years",
"65-74 years",
"75-84 years",
"85 years and older"
),
periodLevels=c("2015-2019", "2020", "2021", "2022"),
periodKeep=c("2015-2019", "2020", "2021"),
yearLevels=2015:2022
)
# STEP 1: Latest CDC all-cause deaths data
cdcLoc <- "Weekly_counts_of_deaths_by_jurisdiction_and_age_group_downloaded_20220623.csv"
cdcList_20220623 <- readRunCDCAllCause(loc=cdcLoc,
weekThru=21,
lst=readFromRDS("cdc_daily_220602"),
stateNoCheck=c(),
pdfCluster=TRUE,
pdfAge=TRUE
)##
## Parameter cvDeathThru has been set as: 2022-05-28
##
##
## *** Data suppression checks ***
##
## Rows in states to be checked that have NA deaths or a note for suppression:
## state weekEnding year week age
## 1 SD 2022-04-30 2022 17 65-74 years
## 2 SD 2022-04-30 2022 17 75-84 years
## Suppress deaths
## 1 Suppressed (counts highly incomplete, <50% of expected) NA
## 2 Suppressed (counts highly incomplete, <50% of expected) NA
##
##
## Problems by state:
## # A tibble: 1 x 5
## noCheck state problem n deaths
## <lgl> <chr> <lgl> <int> <dbl>
## 1 FALSE SD TRUE 2 NA
##
##
## There are 2 rows with errors; maximum for any given state is 2 errors
##
##
## Data suppression checks passed
##
##
## *** File has been checked for uniqueness by: state year week age
##
## Rows: 106,840
## Columns: 12
## $ fullState <chr> "Alabama", "Alabama", "Alabama", "Alabama", "Alabama", "Ala~
## $ weekEnding <date> 2015-01-10, 2015-01-10, 2015-01-10, 2015-01-10, 2015-01-10~
## $ state <chr> "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL",~
## $ year <fct> 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015,~
## $ week <int> 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 4, 4,~
## $ age <fct> Under 25 years, 25-44 years, 45-64 years, 65-74 years, 75-8~
## $ period <fct> 2015-2019, 2015-2019, 2015-2019, 2015-2019, 2015-2019, 2015~
## $ Type <chr> "Predicted (weighted)", "Predicted (weighted)", "Predicted ~
## $ Suppress <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,~
## $ n <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,~
## $ deaths <dbl> 25, 67, 253, 202, 272, 320, 28, 49, 256, 222, 253, 332, 26,~
## $ Note <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,~
##
## Check Control Levels and Record Counts for Processed Data:
##
##
## Checking variable combination: age
## # A tibble: 6 x 4
## age n n_deaths_na deaths
## <fct> <dbl> <dbl> <dbl>
## 1 Under 25 years 12528 0 434501
## 2 25-44 years 16114 0 1115606
## 3 45-64 years 19554 0 4261157
## 4 65-74 years 19547 0 4306424
## 5 75-84 years 19554 0 5271898
## 6 85 years and older 19543 0 6662410
##
##
## Checking variable combination: period year Type
## # A tibble: 8 x 6
## period year Type n n_deaths_na deaths
## <fct> <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-2019 2015 Predicted (weighted) 14367 0 2698242
## 2 2015-2019 2016 Predicted (weighted) 14445 0 2725557
## 3 2015-2019 2017 Predicted (weighted) 14408 0 2802070
## 4 2015-2019 2018 Predicted (weighted) 14400 0 2830373
## 5 2015-2019 2019 Predicted (weighted) 14413 0 2843917
## 6 2020 2020 Predicted (weighted) 14834 0 3432792
## 7 2021 2021 Predicted (weighted) 14698 0 3451431
## 8 2022 2022 Predicted (weighted) 5275 0 1267614
##
##
## Checking variable combination: period Suppress
## # A tibble: 4 x 5
## period Suppress n n_deaths_na deaths
## <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-2019 <NA> 72033 0 13900159
## 2 2020 <NA> 14834 0 3432792
## 3 2021 <NA> 14698 0 3451431
## 4 2022 <NA> 5275 0 1267614
##
##
## Checking variable combination: period Note
## # A tibble: 9 x 5
## period Note n n_deaths_na deaths
## <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-20~ <NA> 72033 0 1.39e7
## 2 2020 Data in recent weeks are incomplete. Only ~ 279 0 8.68e4
## 3 2020 <NA> 14555 0 3.35e6
## 4 2021 Data in recent weeks are incomplete. Only ~ 12116 0 2.42e6
## 5 2021 Data in recent weeks are incomplete. Only ~ 10 0 2.58e2
## 6 2021 Data in recent weeks are incomplete. Only ~ 2572 0 1.04e6
## 7 2022 Data in recent weeks are incomplete. Only ~ 4347 0 1.06e6
## 8 2022 Data in recent weeks are incomplete. Only ~ 76 0 1.80e4
## 9 2022 Data in recent weeks are incomplete. Only ~ 852 0 1.90e5
##
## *** File has been checked for uniqueness by: cluster year week
##
## Plots will be run after excluding stateNoCheck states
##
## Detailed cluster summary PDF file is available at: ./RInputFiles/Coronavirus/Plots/CDC_cluster_2022w21.pdf
##
## Returning plot outputs to the main log file
## Joining, by = "state"
##
## Detailed age summary PDF file is available at: ./RInputFiles/Coronavirus/Plots/CDC_age_2022w21.pdf
##
## Returning plot outputs to the main log file
saveToRDS(cdcList_20220623, ovrWriteError=FALSE)
# STEP 2: Latest death bu location-cause data
allCause_220623 <- analyzeAllCause(loc="COvID_deaths_age_place_20220623.csv",
cdcDailyList=readFromRDS("cdc_daily_220602"),
compareThruDate="2022-05-31"
)## `summarise()` has grouped output by 'State'. You can override using the `.groups` argument.
##
## States without abbreviations
## # A tibble: 2 x 10
## # Groups: State [2]
## State abb Year Month covidDeaths totalDeaths pneumoDeaths pneumoCovidDeat~
## <chr> <chr> <int> <int> <dbl> <dbl> <dbl> <dbl>
## 1 New Y~ <NA> 0 0 35136 170882 22567 13036
## 2 Puert~ <NA> 0 0 4311 78570 11023 3082
## # ... with 2 more variables: fluDeaths <dbl>, pnemoFluCovidDeaths <dbl>
##
## Sub-lists are identical by: asofDate, startDate, endDate, Group, State, deathPlace, Age
## # A tibble: 1,748 x 12
## asofDate startDate endDate Group State deathPlace Age name dfSub
## <date> <date> <date> <chr> <chr> <chr> <chr> <chr> <dbl>
## 1 2022-06-02 2020-10-01 2020-10-31 By Mo~ Unite~ Total - All~ 30-3~ pnem~ 205
## 2 2022-06-02 2021-08-01 2021-08-31 By Mo~ Unite~ Other All ~ pneu~ 671
## 3 2022-06-02 2021-10-01 2021-10-31 By Mo~ Unite~ Decedent's ~ 40-4~ pnem~ 149
## 4 2022-06-02 2020-02-01 2020-02-29 By Mo~ Unite~ Total - All~ 30-3~ pnem~ 71
## 5 2022-06-02 2021-11-01 2021-11-30 By Mo~ Unite~ Healthcare ~ 75-8~ pnem~ 139
## 6 2022-06-02 2020-11-01 2020-11-30 By Mo~ Unite~ Total - All~ 30-3~ pneu~ 227
## 7 2022-06-02 2022-04-01 2022-04-30 By Mo~ Unite~ Total - All~ All ~ fluD~ 168
## 8 2022-06-02 2020-08-01 2020-08-31 By Mo~ Unite~ Other 0-17~ tota~ 116
## 9 2022-06-02 2020-09-01 2020-09-30 By Mo~ Unite~ Decedent's ~ 50-6~ pnem~ 190
## 10 2022-06-02 2021-10-01 2021-10-31 By Mo~ Unite~ Decedent's ~ 65-7~ pneu~ 86
## # ... with 1,738 more rows, and 3 more variables: dfTot <dbl>, delta <dbl>,
## # pct <dbl>
##
## Sub-lists are identical by: asofDate, startDate, endDate, Group, State, deathPlace, Age
## # A tibble: 0 x 12
## # ... with 12 variables: asofDate <date>, startDate <date>, endDate <date>,
## # Group <chr>, State <chr>, deathPlace <chr>, Age <chr>, name <chr>,
## # dfSub <dbl>, dfTot <dbl>, delta <dbl>, pct <dbl>
##
## Sub-lists are identical by: asofDate, startDate, endDate, Group, State, deathPlace, Age
## # A tibble: 0 x 12
## # ... with 12 variables: asofDate <date>, startDate <date>, endDate <date>,
## # Group <chr>, State <chr>, deathPlace <chr>, Age <chr>, name <chr>,
## # dfSub <dbl>, dfTot <dbl>, delta <dbl>, pct <dbl>
## # A tibble: 51 x 4
## abb cumValue tot_deaths pctdiff
## <chr> <dbl> <dbl> <dbl>
## 1 NY 36518 68346 0.304
## 2 DC 2010 1343 0.199
## 3 ND 2777 2283 0.0976
## 4 NC 28931 24660 0.0797
## 5 GA 32614 38198 0.0789
## 6 WY 1577 1820 0.0715
## 7 NE 4947 4290 0.0711
## 8 OH 43659 38628 0.0611
## 9 MI 32215 36357 0.0604
## 10 OK 16139 14420 0.0563
## # ... with 41 more rows
## # A tibble: 1 x 3
## cumValue tot_deaths pctdiff
## <dbl> <dbl> <dbl>
## 1 969868 997512 1.82
saveToRDS(allCause_220623, ovrWriteError=FALSE)
# STEP 3: Facets for excess all-cause deaths
excessDeathFacets(lstCDC=cdcList_20220623, lstAll=allCause_220623, dateThru="2022-04-30", plotYLim=c(-200, 1200))
Updated with the latest data:
# STEP 1: Latest CDC all-cause deaths data
cdcLoc <- "Weekly_counts_of_deaths_by_jurisdiction_and_age_group_downloaded_20220713.csv"
cdcList_20220713 <- readRunCDCAllCause(loc=cdcLoc,
weekThru=24,
lst=readFromRDS("cdc_daily_220704"),
stateNoCheck=c(),
pdfCluster=TRUE,
pdfAge=TRUE
)##
## Parameter cvDeathThru has been set as: 2022-06-18
##
##
## *** Data suppression checks ***
##
## Rows in states to be checked that have NA deaths or a note for suppression:
## [1] state weekEnding year week age Suppress deaths
## <0 rows> (or 0-length row.names)
##
##
## Problems by state:
## # A tibble: 0 x 5
## # ... with 5 variables: noCheck <lgl>, state <chr>, problem <lgl>, n <int>,
## # deaths <dbl>## Warning in max(.): no non-missing arguments to max; returning -Inf##
##
## There are 0 rows with errors; maximum for any given state is -Inf errors
##
##
## Data suppression checks passed
##
##
## *** File has been checked for uniqueness by: state year week age
##
## Rows: 108,099
## Columns: 12
## $ fullState <chr> "Alabama", "Alabama", "Alabama", "Alabama", "Alabama", "Ala~
## $ weekEnding <date> 2015-01-10, 2015-01-10, 2015-01-10, 2015-01-10, 2015-01-10~
## $ state <chr> "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL",~
## $ year <fct> 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015,~
## $ week <int> 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 4, 4,~
## $ age <fct> Under 25 years, 25-44 years, 45-64 years, 65-74 years, 75-8~
## $ period <fct> 2015-2019, 2015-2019, 2015-2019, 2015-2019, 2015-2019, 2015~
## $ Type <chr> "Predicted (weighted)", "Predicted (weighted)", "Predicted ~
## $ Suppress <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,~
## $ n <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,~
## $ deaths <dbl> 25, 67, 253, 202, 272, 320, 28, 49, 256, 222, 253, 332, 26,~
## $ Note <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,~
##
## Check Control Levels and Record Counts for Processed Data:
##
##
## Checking variable combination: age
## # A tibble: 6 x 4
## age n n_deaths_na deaths
## <fct> <dbl> <dbl> <dbl>
## 1 Under 25 years 12543 0 432096
## 2 25-44 years 16323 0 1118247
## 3 45-64 years 19812 0 4307809
## 4 65-74 years 19806 0 4368517
## 5 75-84 years 19813 0 5351113
## 6 85 years and older 19802 0 6752462
##
##
## Checking variable combination: period year Type
## # A tibble: 8 x 6
## period year Type n n_deaths_na deaths
## <fct> <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-2019 2015 Predicted (weighted) 14367 0 2698242
## 2 2015-2019 2016 Predicted (weighted) 14445 0 2725557
## 3 2015-2019 2017 Predicted (weighted) 14408 0 2802070
## 4 2015-2019 2018 Predicted (weighted) 14400 0 2830373
## 5 2015-2019 2019 Predicted (weighted) 14413 0 2843917
## 6 2020 2020 Predicted (weighted) 14834 0 3432816
## 7 2021 2021 Predicted (weighted) 14702 0 3450646
## 8 2022 2022 Predicted (weighted) 6530 0 1546623
##
##
## Checking variable combination: period Suppress
## # A tibble: 4 x 5
## period Suppress n n_deaths_na deaths
## <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-2019 <NA> 72033 0 13900159
## 2 2020 <NA> 14834 0 3432816
## 3 2021 <NA> 14702 0 3450646
## 4 2022 <NA> 6530 0 1546623
##
##
## Checking variable combination: period Note
## # A tibble: 9 x 5
## period Note n n_deaths_na deaths
## <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-20~ <NA> 72033 0 1.39e7
## 2 2020 Data in recent weeks are incomplete. Only ~ 279 0 8.69e4
## 3 2020 <NA> 14555 0 3.35e6
## 4 2021 Data in recent weeks are incomplete. Only ~ 13990 0 3.20e6
## 5 2021 Data in recent weeks are incomplete. Only ~ 15 0 4.01e2
## 6 2021 Data in recent weeks are incomplete. Only ~ 697 0 2.51e5
## 7 2022 Data in recent weeks are incomplete. Only ~ 1058 0 1.61e5
## 8 2022 Data in recent weeks are incomplete. Only ~ 86 0 7.94e3
## 9 2022 Data in recent weeks are incomplete. Only ~ 5386 0 1.38e6
##
## *** File has been checked for uniqueness by: cluster year week
##
## Plots will be run after excluding stateNoCheck states
##
## Detailed cluster summary PDF file is available at: ./RInputFiles/Coronavirus/Plots/CDC_cluster_2022w24.pdf
##
## Returning plot outputs to the main log file
## Joining, by = "state"
##
## Detailed age summary PDF file is available at: ./RInputFiles/Coronavirus/Plots/CDC_age_2022w24.pdf
##
## Returning plot outputs to the main log file
saveToRDS(cdcList_20220713, ovrWriteError=FALSE)
# STEP 2: Latest death bu location-cause data
allCause_220713 <- analyzeAllCause(loc="COvID_deaths_age_place_20220713.csv",
cdcDailyList=readFromRDS("cdc_daily_220704"),
compareThruDate="2022-06-30"
)## `summarise()` has grouped output by 'State'. You can override using the `.groups` argument.
##
## States without abbreviations
## # A tibble: 2 x 10
## # Groups: State [2]
## State abb Year Month covidDeaths totalDeaths pneumoDeaths pneumoCovidDeat~
## <chr> <chr> <int> <int> <dbl> <dbl> <dbl> <dbl>
## 1 New Y~ <NA> 0 0 35270 174129 22877 13064
## 2 Puert~ <NA> 0 0 4459 80624 11310 3179
## # ... with 2 more variables: fluDeaths <dbl>, pnemoFluCovidDeaths <dbl>
##
## Sub-lists are identical by: asofDate, startDate, endDate, Group, State, deathPlace, Age
## # A tibble: 1,818 x 12
## asofDate startDate endDate Group State deathPlace Age name dfSub
## <date> <date> <date> <chr> <chr> <chr> <chr> <chr> <dbl>
## 1 2022-07-06 2020-10-01 2020-10-31 By Mo~ Unite~ Total - All~ 30-3~ pnem~ 205
## 2 2022-07-06 2021-10-01 2021-10-31 By Mo~ Unite~ Decedent's ~ 40-4~ pnem~ 150
## 3 2022-07-06 2020-02-01 2020-02-29 By Mo~ Unite~ Total - All~ 30-3~ pnem~ 71
## 4 2022-07-06 2021-11-01 2021-11-30 By Mo~ Unite~ Healthcare ~ 75-8~ pnem~ 139
## 5 2022-07-06 2022-04-01 2022-04-30 By Mo~ Unite~ Total - All~ All ~ fluD~ 184
## 6 2022-07-06 2020-11-01 2020-11-30 By Mo~ Unite~ Total - All~ 30-3~ pneu~ 227
## 7 2022-07-06 2021-08-01 2021-08-31 By Mo~ Unite~ Other All ~ pneu~ 627
## 8 2022-07-06 2022-06-01 2022-06-30 By Mo~ Unite~ Decedent's ~ 85 y~ pneu~ 183
## 9 2022-07-06 2020-01-01 2022-07-02 By To~ Unite~ Total - All~ 0-17~ fluD~ 50
## 10 2022-07-06 2020-01-01 2022-07-02 By To~ Unite~ Total - All~ 30-3~ fluD~ 200
## # ... with 1,808 more rows, and 3 more variables: dfTot <dbl>, delta <dbl>,
## # pct <dbl>
##
## Sub-lists are identical by: asofDate, startDate, endDate, Group, State, deathPlace, Age
## # A tibble: 0 x 12
## # ... with 12 variables: asofDate <date>, startDate <date>, endDate <date>,
## # Group <chr>, State <chr>, deathPlace <chr>, Age <chr>, name <chr>,
## # dfSub <dbl>, dfTot <dbl>, delta <dbl>, pct <dbl>
##
## Sub-lists are identical by: asofDate, startDate, endDate, Group, State, deathPlace, Age
## # A tibble: 0 x 12
## # ... with 12 variables: asofDate <date>, startDate <date>, endDate <date>,
## # Group <chr>, State <chr>, deathPlace <chr>, Age <chr>, name <chr>,
## # dfSub <dbl>, dfTot <dbl>, delta <dbl>, pct <dbl>
## # A tibble: 51 x 4
## abb cumValue tot_deaths pctdiff
## <chr> <dbl> <dbl> <dbl>
## 1 NY 36925 69007 0.303
## 2 DC 1994 1351 0.192
## 3 WY 1462 1834 0.113
## 4 ND 2802 2296 0.0993
## 5 GA 32661 38579 0.0831
## 6 NC 29438 25211 0.0773
## 7 MI 32104 36918 0.0697
## 8 NE 4986 4342 0.0690
## 9 AZ 26808 30515 0.0647
## 10 OH 44034 38852 0.0625
## # ... with 41 more rows
## # A tibble: 1 x 3
## cumValue tot_deaths pctdiff
## <dbl> <dbl> <dbl>
## 1 974598 1008140 1.91
## Warning: Removed 8 rows containing missing values (geom_col).
## Warning: Removed 8 rows containing missing values (geom_col).
saveToRDS(allCause_220713, ovrWriteError=FALSE)
# STEP 3: Facets for excess all-cause deaths
excessDeathFacets(lstCDC=cdcList_20220713, lstAll=allCause_220713, dateThru="2022-05-31", plotYLim=c(-200, 1200))
There have been issues with US all-cause deaths data since a “systems upgrade” in mid-June. How much restatement of data has occurred?
# Mapping file of epiweek and epiyear to date
mapEpi <- tibble::tibble(date=seq.Date(as.Date("2014-12-01"), as.Date("2031-01-31"), by=1)) %>%
mutate(epiYear=as.integer(lubridate::epiyear(date)), epiWeek=as.integer(lubridate::epiweek(date)))
nameFile <- "ageAgg"
dfCheck <- bind_rows(readFromRDS("cdcList_20220713")[[nameFile]],
readFromRDS("cdcList_20220623")[[nameFile]],
readFromRDS("cdcList_20220105")[[nameFile]],
.id="fileDate"
) %>%
mutate(fileDate=c("1"="2022-07-13", "2"="2022-06-23", "3"="2022-01-05")[fileDate])
mapEpi %>%
arrange(date) %>%
group_by(epiYear, epiWeek) %>%
filter(row_number()==1) %>%
ungroup() %>%
rename(yearint=epiYear, week=epiWeek) %>%
right_join(dfCheck, by=c("yearint", "week")) %>%
ggplot(aes(x=date, y=deaths)) +
geom_line(aes(color=fileDate, group=fileDate)) +
lims(y=c(0, NA)) +
labs(x=NULL, y="Reported all-cause US deaths", title="US all-cause deaths by report date") +
facet_wrap(~age, scales="free_y")
Data appear anomalous, particularly 2022 deaths in “Under 25 years” and “25-44 years”. Partly, this is incomplete reporting in the most recent weeks (normal), but partly this may be driven by data not yet re-entered after the upgrade. It is striking that there are fewer reported all-cause deaths in the 2022-07-13 data than in the 2022-06-23 data for any cohort, as all-cause data almost always increases as additional reports are received from vital statistics departments. Trends among “45-64 years” and senior citizens, at a glance, are the more commonly observed build over time
The process is converted to functional form:
makeRestatementData <- function(vecFiles, key, vecNames=NULL, epiRange=as.Date(c("2014-12-01", "2031-01-31"))) {
# FUNCTION ARGUMENTS:
# vecFiles: character vector of file names (will be extracted using readFromRDS)
# key: the extract element from each of the lists
# vecNames: names to be used in plot for each of the extracts (NULL means infer from ...)
# epiRange: range for converting epiweek and epiyear to date (should be a larger range than data)
# Add names to vecNames if not passed
if(!is.null(vecNames) & is.null(names(vecNames)))
vecNames <- vecNames %>% purrr::set_names(as.character(1:length(vecFiles)))
# Create keyNames if not provided
if(is.null(vecNames)) {
vecNames <- as.character(lubridate::ymd(stringr::str_remove(vecFiles, ".*_"))) %>%
purrr::set_names(as.character(1:length(vecFiles)))
}
# Create epi mapping file
dfEpi <- tibble::tibble(date=seq.Date(epiRange[1], epiRange[2], by=1)) %>%
mutate(epiYear=as.integer(lubridate::epiyear(date)),
epiWeek=as.integer(lubridate::epiweek(date))
)
# Create single date for each epiWeek and epiYear
mapEpi <- dfEpi %>%
arrange(date) %>%
group_by(epiYear, epiWeek) %>%
filter(row_number()==1) %>%
ungroup() %>%
rename(yearint=epiYear, week=epiWeek)
# Read and integrate file, add epiDate
purrr::map_dfr(.x=vecFiles,
.f=function(x) readFromRDS(x)[[key]],
.id="fileDate"
) %>%
mutate(fileDate=vecNames[fileDate]) %>%
left_join(mapEpi, by=c("yearint", "week"))
}
plotRestatementData <- function(df, wrapBy=NULL, asRatio=FALSE) {
# FUNCTION ARGUMENTS:
# df: data frame or tibble formatted for plotting
# wrapBy: variable for facet_wrap (NULL means infer from file, FALSE means do not wrap)
# asRatio: boolean, should ratios be plotted rather than values?
# Create the appropriate wrapBy if passed as NULL
if(is.null(wrapBy)) {
if("age" %in% names(df)) wrapBy <- "age"
else if ("state" %in% names(df)) wrapBy <- "state"
else if ("cluster" %in% names(df)) wrapBy <- "cluster"
else wrapBy <- FALSE
}
plotTitle <- "US all-cause deaths by report date"
plotSubTitle <- NULL
plotYAxis <- "Reported all-cause US deaths"
# Create ratios if appropriate
if(isTRUE(asRatio)) {
groupVars <- c("date")
if(!isFALSE(wrapBy)) groupVars <- c(groupVars, wrapBy)
df <- df %>%
rename(trueFileDate=fileDate, trueDeaths=deaths) %>%
arrange(trueFileDate) %>%
group_by_at(all_of(groupVars)) %>%
mutate(n=n(),
fileDate=ifelse(row_number()==1, trueFileDate, paste0(trueFileDate, " vs. ", lag(trueFileDate))),
deaths=ifelse(row_number()==1, trueDeaths, trueDeaths/lag(trueDeaths))
) %>%
ungroup()
plotTitle <- "Ratio of US all-cause deaths by report date"
plotSubTitle <- "Ratios filtered to exclude NA and results greater than 3"
plotYAxis <- "Ratio of reported all-cause US deaths"
}
# Create base plot
p1 <- df %>%
filter(if(isTRUE(asRatio)) fileDate != min(fileDate) else TRUE) %>%
filter(if(isTRUE(asRatio)) !is.na(deaths) & deaths <= 3 else TRUE) %>%
ggplot(aes(x=date, y=deaths)) +
geom_line(aes(color=fileDate, group=fileDate)) +
lims(y=c(0, NA)) +
labs(x=NULL, y=plotYAxis, subtitle=plotSubTitle, title=plotTitle) +
scale_color_discrete("File Date")
# Add line at 1.0 if ratio
if(isTRUE(asRatio)) p1 <- p1 + geom_hline(yintercept=1, lty=2)
# Add facetting if appropriate
if(!isFALSE(wrapBy)) p1 <- p1 + facet_wrap(~get(wrapBy), scales="free_y")
# Print the plot
print(p1)
}
makeRestatementData(c("cdcList_20220713", "cdcList_20220623", "cdcList_20220105"), key="ageAgg")## # A tibble: 6,810 x 12
## fileDate age year week deaths weekfct yearint pred delta cumDelta
## <chr> <fct> <fct> <int> <dbl> <fct> <int> <dbl> <dbl> <dbl>
## 1 2022-07-13 Under 25 ~ 2015 1 1069 1 2015 1143. -74.4 -74.4
## 2 2022-07-13 Under 25 ~ 2016 1 1067 1 2016 1122. -55.0 -55.0
## 3 2022-07-13 Under 25 ~ 2017 1 1147 1 2017 1101. 46.4 46.4
## 4 2022-07-13 Under 25 ~ 2018 1 1185 1 2018 1079. 106. 106.
## 5 2022-07-13 Under 25 ~ 2019 1 1035 1 2019 1058. -22.8 -22.8
## 6 2022-07-13 Under 25 ~ 2020 1 1101 1 2020 1036. 64.6 64.6
## 7 2022-07-13 Under 25 ~ 2021 1 1072 1 2021 1015. 57.0 57.0
## 8 2022-07-13 Under 25 ~ 2022 1 931 1 2022 994. -62.6 -62.6
## 9 2022-07-13 Under 25 ~ 2015 2 1103 2 2015 1133. -30.0 -104.
## 10 2022-07-13 Under 25 ~ 2016 2 1068 2 2016 1112. -43.6 -98.6
## # ... with 6,800 more rows, and 2 more variables: cumPred <dbl>, date <date>makeRestatementData(c("cdcList_20220713", "cdcList_20220623", "cdcList_20220105"), key="ageAgg") %>%
plotRestatementData()
makeRestatementData(c("cdcList_20220713", "cdcList_20220623", "cdcList_20220105"), key="ageAgg") %>%
plotRestatementData(asRatio=TRUE)
makeRestatementData(c("cdcList_20220713", "cdcList_20220623", "cdcList_20220105"), key="allUSAgg") %>%
plotRestatementData()
makeRestatementData(c("cdcList_20220713", "cdcList_20220623", "cdcList_20220105"), key="allUSAgg") %>%
plotRestatementData(asRatio=TRUE)
Github user USMortality stores archived all-cause deaths data. The file from 2022 week 17 is downloaded and processed:
# STEP 1: Archived CDC all-cause deaths data
cdcLoc <- "Weekly_counts_of_deaths_by_jurisdiction_and_age_group_2022_17.txt"
cdcList_arch_2022w17 <- readRunCDCAllCause(loc=cdcLoc,
weekThru=16,
lst=readFromRDS("cdc_daily_220704"),
stateNoCheck=c(),
pdfCluster=TRUE,
pdfAge=TRUE
)##
## Parameter cvDeathThru has been set as: 2022-04-23
##
##
## *** Data suppression checks ***
##
## Rows in states to be checked that have NA deaths or a note for suppression:
## state weekEnding year week age
## 1 NE 2022-04-23 2022 16 65-74 years
## 2 NE 2022-04-23 2022 16 75-84 years
## 3 NE 2022-04-23 2022 16 85 years and older
## 4 IN 2022-04-16 2022 15 25-44 years
## 5 IN 2022-04-16 2022 15 45-64 years
## 6 IN 2022-04-16 2022 15 65-74 years
## 7 IN 2022-04-16 2022 15 75-84 years
## 8 IN 2022-04-16 2022 15 85 years and older
## Suppress deaths
## 1 Suppressed (counts highly incomplete, <50% of expected) NA
## 2 Suppressed (counts highly incomplete, <50% of expected) NA
## 3 Suppressed (counts highly incomplete, <50% of expected) NA
## 4 Suppressed (counts highly incomplete, <50% of expected) NA
## 5 Suppressed (counts highly incomplete, <50% of expected) NA
## 6 Suppressed (counts highly incomplete, <50% of expected) NA
## 7 Suppressed (counts highly incomplete, <50% of expected) NA
## 8 Suppressed (counts highly incomplete, <50% of expected) NA
##
##
## Problems by state:
## # A tibble: 2 x 5
## noCheck state problem n deaths
## <lgl> <chr> <lgl> <int> <dbl>
## 1 FALSE IN TRUE 5 NA
## 2 FALSE NE TRUE 3 NA
##
##
## There are 8 rows with errors; maximum for any given state is 5 errors
##
##
## Data suppression checks passed
##
##
## *** File has been checked for uniqueness by: state year week age
##
## Rows: 105,996
## Columns: 12
## $ fullState <chr> "Alabama", "Alabama", "Alabama", "Alabama", "Alabama", "Ala~
## $ weekEnding <date> 2015-01-10, 2015-01-10, 2015-01-10, 2015-01-10, 2015-01-10~
## $ state <chr> "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL", "AL",~
## $ year <fct> 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015,~
## $ week <int> 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 4, 4,~
## $ age <fct> Under 25 years, 25-44 years, 45-64 years, 65-74 years, 75-8~
## $ period <fct> 2015-2019, 2015-2019, 2015-2019, 2015-2019, 2015-2019, 2015~
## $ Type <chr> "Predicted (weighted)", "Predicted (weighted)", "Predicted ~
## $ Suppress <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,~
## $ n <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,~
## $ deaths <dbl> 25, 67, 253, 202, 272, 320, 28, 49, 256, 222, 253, 332, 26,~
## $ Note <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,~
##
## Check Control Levels and Record Counts for Processed Data:
##
##
## Checking variable combination: age
## # A tibble: 6 x 4
## age n n_deaths_na deaths
## <fct> <dbl> <dbl> <dbl>
## 1 Under 25 years 12422 0 430722
## 2 25-44 years 15982 0 1105179
## 3 45-64 years 19401 0 4228337
## 4 65-74 years 19397 0 4270304
## 5 75-84 years 19403 0 5227671
## 6 85 years and older 19391 0 6612949
##
##
## Checking variable combination: period year Type
## # A tibble: 8 x 6
## period year Type n n_deaths_na deaths
## <fct> <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-2019 2015 Predicted (weighted) 14367 0 2698242
## 2 2015-2019 2016 Predicted (weighted) 14445 0 2725557
## 3 2015-2019 2017 Predicted (weighted) 14408 0 2802070
## 4 2015-2019 2018 Predicted (weighted) 14400 0 2830373
## 5 2015-2019 2019 Predicted (weighted) 14413 0 2843917
## 6 2020 2020 Predicted (weighted) 14834 0 3432787
## 7 2021 2021 Predicted (weighted) 14696 0 3452019
## 8 2022 2022 Predicted (weighted) 4433 0 1090197
##
##
## Checking variable combination: period Suppress
## # A tibble: 4 x 5
## period Suppress n n_deaths_na deaths
## <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-2019 <NA> 72033 0 13900159
## 2 2020 <NA> 14834 0 3432787
## 3 2021 <NA> 14696 0 3452019
## 4 2022 <NA> 4433 0 1090197
##
##
## Checking variable combination: period Note
## # A tibble: 8 x 5
## period Note n n_deaths_na deaths
## <fct> <chr> <dbl> <dbl> <dbl>
## 1 2015-20~ <NA> 72033 0 1.39e7
## 2 2020 Data in recent weeks are incomplete. Only ~ 279 0 8.68e4
## 3 2020 <NA> 14555 0 3.35e6
## 4 2021 Data in recent weeks are incomplete. Only ~ 12124 0 2.39e6
## 5 2021 Data in recent weeks are incomplete. Only ~ 2572 0 1.06e6
## 6 2022 Data in recent weeks are incomplete. Only ~ 3310 0 8.36e5
## 7 2022 Data in recent weeks are incomplete. Only ~ 77 0 1.76e4
## 8 2022 Data in recent weeks are incomplete. Only ~ 1046 0 2.37e5
##
## *** File has been checked for uniqueness by: cluster year week
##
## Plots will be run after excluding stateNoCheck states
##
## Detailed cluster summary PDF file is available at: ./RInputFiles/Coronavirus/Plots/CDC_cluster_2022w16.pdf
##
## Returning plot outputs to the main log file
## Joining, by = "state"
##
## Detailed age summary PDF file is available at: ./RInputFiles/Coronavirus/Plots/CDC_age_2022w16.pdf
##
## Returning plot outputs to the main log file
saveToRDS(cdcList_arch_2022w17, ovrWriteError=FALSE)Comparisons can be run among deaths in each dataset:
makeRestatementData(c("cdcList_20220713", "cdcList_arch_2022w17", "cdcList_20220105"),
key="allUSAgg",
vecNames=c("2022-07-13", "2022-04-25", "2022-01-05")
) %>%
plotRestatementData(asRatio=TRUE)
makeRestatementData(c("cdcList_20220713", "cdcList_arch_2022w17", "cdcList_20220105"),
key="ageAgg",
vecNames=c("2022-07-13", "2022-04-25", "2022-01-05")
) %>%
plotRestatementData(asRatio=TRUE)
makeRestatementData(c("cdcList_20220713", "cdcList_arch_2022w17", "cdcList_20220105"),
key="clusterAgg",
vecNames=c("2022-07-13", "2022-04-25", "2022-01-05")
) %>%
plotRestatementData(asRatio=TRUE)
The persistent gap between reported deaths in 2022-01-03 and later reports is the exclusion of several cluster 5 states from the 2022-01-03 processing due to data suppression issues. There continues to be an anomaly where deaths among people under age 45 decreased between 2022-04-25 and 2022-07-13. This trend of decreasing deaths is significantly reduced or not existent in data for ages 45+
Data prior to exclusions are examined for consistency:
dfCheck <- readFromRDS("cdcList_arch_2022w17")$cdc %>%
select(state, weekEnding, age, deaths_220425=deaths) %>%
full_join(readFromRDS("cdcList_20220713")$cdc %>% select(state, weekEnding, age, deaths_220713=deaths),
by=c("state", "weekEnding", "age")
) %>%
mutate(delta=ifelse(is.na(deaths_220713), 0, deaths_220713)-ifelse(is.na(deaths_220425), 0, deaths_220425),
neg=(delta < 0)
)
dfCheck## # A tibble: 108,190 x 7
## state weekEnding age deaths_220425 deaths_220713 delta neg
## <chr> <date> <fct> <dbl> <dbl> <dbl> <lgl>
## 1 AL 2015-01-10 Under 25 years 25 25 0 FALSE
## 2 AL 2015-01-10 25-44 years 67 67 0 FALSE
## 3 AL 2015-01-10 45-64 years 253 253 0 FALSE
## 4 AL 2015-01-10 65-74 years 202 202 0 FALSE
## 5 AL 2015-01-10 75-84 years 272 272 0 FALSE
## 6 AL 2015-01-10 85 years and older 320 320 0 FALSE
## 7 AL 2015-01-17 Under 25 years 28 28 0 FALSE
## 8 AL 2015-01-17 25-44 years 49 49 0 FALSE
## 9 AL 2015-01-17 45-64 years 256 256 0 FALSE
## 10 AL 2015-01-17 65-74 years 222 222 0 FALSE
## # ... with 108,180 more rowsdfCheck %>% count(neg)## # A tibble: 2 x 2
## neg n
## <lgl> <int>
## 1 FALSE 104486
## 2 TRUE 3704# Get counts of changes by state
dfCheck %>%
group_by(state) %>%
summarize(nNeg=sum(neg), negDelta=sum(delta*neg), n=n(), .groups="drop") %>%
ggplot(aes(x=fct_reorder(state, negDelta), y=negDelta)) +
geom_col(fill="lightblue") +
geom_text(aes(label=negDelta), hjust=1) +
coord_flip() +
labs(y="Sum of negative changes in weekly deaths by age group from 2022-04-25 to 2022-07-13",
x=NULL,
title="Negative change in weekly death by state summary"
)
# Examples overall
dfCheck %>%
select(-delta, -neg) %>%
pivot_longer(starts_with("deaths")) %>%
group_by(weekEnding, age, name) %>%
summarize(deaths=specNA()(value), .groups="drop") %>%
ggplot(aes(x=weekEnding, y=deaths)) +
geom_line(aes(group=name, color=name)) +
lims(y=c(0, NA)) +
labs(x=NULL, y="Reported deaths", title="Reported deaths by age group and week in US") +
facet_wrap(~age, scales="free_y")## Warning: Removed 8 row(s) containing missing values (geom_path).
# Examples from Florida (biggest change)
dfCheck %>%
filter(state=="FL") %>%
select(-delta, -neg) %>%
pivot_longer(starts_with("deaths")) %>%
ggplot(aes(x=weekEnding, y=value)) +
geom_line(aes(group=name, color=name)) +
lims(y=c(0, NA)) +
labs(x=NULL, y="Reported deaths", title="Reported deaths by age group and week in Florida") +
facet_wrap(~age, scales="free_y")## Warning: Removed 8 row(s) containing missing values (geom_path).
Florida data shows similarities to the national data, with negative restatements and negative recent trends primarily limited to the 0-44 years buckets.
Each state and age group is assessed for the total amount of negative delta relative to the average number of annual deaths in the group:
dfCheckAvg <- dfCheck %>%
group_by(state, age) %>%
summarize(across(starts_with("deaths"), specNA(mean)),
delta=specNA(sum)(ifelse(neg, delta, 0)),
.groups="drop"
) %>%
mutate(deltaRatio=delta/deaths_220425)
dfCheckAvg %>%
ggplot(aes(x=fct_reorder(state, deltaRatio, min), y=deltaRatio)) +
geom_col(fill="lightblue") +
geom_text(aes(y=deltaRatio/2, label=round(deltaRatio, 1))) +
coord_flip() +
facet_wrap(~age, nrow=1) +
labs(title="Total negative restatement", subtitle="Units are average number of weeks", y="Avg weeks", x=NULL)
bigDelta <- c("CO", "AZ", "SC", "FL", "OK", "VT")
dfCheck %>%
mutate(type=ifelse(state %in% bigDelta, "big delta", "all other")) %>%
group_by(weekEnding, type, age) %>%
summarize(across(starts_with("deaths"), specNA(sum)), .groups="drop") %>%
mutate(daynum=1L+7*as.integer(weekEnding-min(weekEnding))) %>%
mutate(pred=predict(lm(deaths_220425 ~ daynum*type*age, data=., subset=lubridate::year(weekEnding)<=2019),
newdata=.
)
) %>%
select(-daynum) %>%
pivot_longer(-c(weekEnding, type, age, pred)) %>%
ggplot(aes(x=weekEnding, y=value)) +
geom_line(aes(group=name, color=name)) +
geom_line(aes(y=pred), lty=2, lwd=0.5) +
lims(y=c(0, NA)) +
labs(title="Weekly deaths by state type",
subtitle="Big delta states: CO, AZ, SC, FL, OK, VT\nDashed line is simple linear model using 2015-2019 data",
x=NULL,
y=NULL
) +
facet_grid(type~age, scales="free_y")## Warning: Removed 8 row(s) containing missing values (geom_path).
Much of the negative restatement is driven by a handful of states. There remains a general pattern of deaths, especially among younger groups, falling below historical trends in the most recent data
Plots are created as ratios vs. expected (trend from 2015-2019):
dfCheck %>%
mutate(type=ifelse(state %in% bigDelta, "big delta", "all other")) %>%
group_by(weekEnding, type, age) %>%
summarize(across(starts_with("deaths"), specNA(sum)), .groups="drop") %>%
mutate(daynum=1L+7*as.integer(weekEnding-min(weekEnding))) %>%
mutate(pred=predict(lm(deaths_220425 ~ daynum*type*age, data=., subset=lubridate::year(weekEnding)<=2019),
newdata=.
)
) %>%
select(-daynum) %>%
pivot_longer(-c(weekEnding, type, age, pred)) %>%
ggplot(aes(x=weekEnding, y=value/pred)) +
geom_line(aes(group=name, color=name)) +
geom_line(aes(y=1), lty=2, lwd=0.5) +
lims(y=c(0, NA)) +
labs(title="Ratio of weekly deaths vs. 2015-2019 trend by state type",
subtitle="Big delta states: CO, AZ, SC, FL, OK, VT",
x=NULL,
y=NULL
) +
facet_grid(type~age, scales="free_y")## Warning: Removed 8 row(s) containing missing values (geom_path).
Reported deaths in recent weeks in the “Under 25 years” bucket are under 50% of trends using a simple linear model on 2015-2019 data. The “25-44 years” bucket is ~25% under trend, while the remaining buckets are near trend.
The process is converted to functional form:
calculateRestatementFromRaw <- function(lst1Name,
lst2Name,
lstLabels,
labelBase=FALSE
) {
# FUNCTION ARGUMENTS
# lst1Name: character name (for readFromRDS) of first list that includes raw CDC data
# lst2Name: character name (for readFromRDS) of second list that includes raw CDC data
# lstLabels: labels to be used for list data (e.g., c("deaths_220425", "deaths_220713"))
# labelBase: boolean, should a convenience column "base" be created from lst1Name for later summarization?
# Create the data
df <- readFromRDS(lst1Name)$cdc %>%
select(state, weekEnding, age, deaths) %>%
colRenamer(c("deaths"="deaths1")) %>%
full_join(readFromRDS(lst2Name)$cdc %>%
select(state, weekEnding, age, deaths) %>%
colRenamer(c("deaths"="deaths2")),
by=c("state", "weekEnding", "age")
) %>%
mutate(delta=ifelse(is.na(deaths2), 0, deaths2)-ifelse(is.na(deaths1), 0, deaths1),
neg=(delta < 0)
)
# Add the base column if requested
if(isTRUE(labelBase)) df <- df %>% mutate(base=deaths1)
# Rename and return the data
df %>%
colRenamer(c("deaths1"=lstLabels[1], "deaths2"=lstLabels[2]))
}
identical(dfCheck,
calculateRestatementFromRaw("cdcList_arch_2022w17",
"cdcList_20220713",
lstLabels=c("deaths_220425", "deaths_220713")
)
)## [1] TRUEplotRestatementFromRaw <- function(df,
varNegTotal=c(),
fnDateStack=NULL,
timePeriod=NULL,
makeProp=FALSE,
makePropYears=NULL
) {
# FUNCTION ARGUMENTS:
# df: data frame from calculateRestatementFromRaw
# varNegTotal: variables that should be plotted for sum of negative restatement
# fnDateStack: function to apply to weekEnding for stacking data (NULL means none)
# timePeriod: character vector of time period of two data sources (NULL means infer from variable names)
# makeProp: boolean, should proportional deaths be shown?
# makePropYears: integer vector of years to include in proportional chart (NULL means latest year in data)
if(is.null(timePeriod)) {
timePeriod <- df %>%
select(starts_with("deaths_")) %>%
names() %>%
str_remove(pattern="deaths_") %>%
lubridate::ymd() %>%
as.character() %>%
paste0(collapse=" data to ")
timePeriod <- paste0("from ", timePeriod, " data")
}
# Get counts of changes by varNegTotal
for (keyVar in varNegTotal) {
# Set up data for stacking
if(!is.null(fnDateStack)) {
dfPlot <- df %>%
mutate(stackVar=fnDateStack(weekEnding))
keyVar <- c(keyVar, "stackVar")
} else {
dfPlot <- df
}
# Create the totals
dfTot <- dfPlot %>%
group_by_at(all_of(keyVar[keyVar != "stackVar"])) %>%
summarize(negDelta=sum(delta*neg), .groups="drop")
# Set up base plot and labels
p1 <- dfPlot %>%
group_by_at(all_of(keyVar)) %>%
summarize(nNeg=sum(neg), negDelta=sum(delta*neg), n=n(), .groups="drop") %>%
ggplot(aes(x=fct_reorder(get(keyVar[keyVar != "stackVar"]), negDelta), y=negDelta)) +
geom_text(data=dfTot, aes(label=negDelta), hjust=1) +
coord_flip() +
labs(y=paste0("Sum of negative changes in weekly deaths ", timePeriod),
x=NULL,
title=paste0("Negative change in weekly death by ", keyVar[keyVar != "stackVar"])
)
# Add the columns (either basic or stacked)
if(is.null(fnDateStack)) p1 <- p1 + geom_col(fill="lightblue")
else p1 <- p1 + geom_col(aes(fill=stackVar), position="stack")
# Print the plot
print(p1)
# Create proportional plot if requested
if(isTRUE(makeProp)) {
# Get the year if passed as NULL
if(is.null(makePropYears)) makePropYears <- max(lubridate::year(dfPlot$weekEnding))
# Create the plot
p2 <- dfPlot %>%
filter(lubridate::year(weekEnding) %in% all_of(makePropYears)) %>%
mutate(deltaNeg=ifelse(neg, delta, 0)) %>%
group_by_at(all_of(keyVar[keyVar != "stackVar"])) %>%
summarize(across(where(is.numeric), sum, na.rm=TRUE)) %>%
mutate(pctNeg=deltaNeg/base) %>%
ggplot(aes(x=fct_reorder(get(keyVar[keyVar != "stackVar"]), pctNeg), y=pctNeg)) +
geom_col(fill="lightblue") +
geom_text(aes(y=pctNeg/2, label=paste0(round(100*pctNeg, 1), "%"))) +
coord_flip() +
labs(title=paste0("Proportion of ",
paste0(makePropYears, collapse="-"),
" deaths negatively restated ",
timePeriod
),
y=NULL,
x=NULL
)
# Print the plot
print(p2)
}
}
}
calculateRestatementFromRaw("cdcList_20220105",
"cdcList_arch_2022w17",
lstLabels=c("deaths_220105", "deaths_220425"),
labelBase=TRUE
) %>%
plotRestatementFromRaw(varNegTotal=c("state", "age"), makeProp=TRUE, makePropYears=2021)
calculateRestatementFromRaw("cdcList_arch_2022w17",
"cdcList_20220713",
lstLabels=c("deaths_220425", "deaths_220713"),
labelBase=TRUE
) %>%
plotRestatementFromRaw(varNegTotal=c("state", "age"), makeProp=TRUE, makePropYears=2022)
# Create function for custom quarter-year
tmpCustomQuarter <- function(x)
ifelse(lubridate::year(x)==2022, paste0(lubridate::year(x), "-Q", lubridate::quarter(x)), lubridate::year(x))
calculateRestatementFromRaw("cdcList_arch_2022w17",
"cdcList_20220713",
lstLabels=c("deaths_220425", "deaths_220713")
) %>%
plotRestatementFromRaw(varNegTotal=c("state", "age"),
fnDateStack=tmpCustomQuarter
)
Between the 2022-01-05 data and the 2022-04-25 data, negative restatements were 559 (104+455) among people under the age of 45. Between the 2022-04-25 data and the 2022-07-13 data, negative restatements were 11,236 (3,820 + 7,416) among people under the age of 45. The majority of the 2022-07-13 vs 2022-04-25 restatements are in 2022-Q1 data, and proportionally the younger population is much more heavily restated than the older population
Functions are run on data from previous years:
calculateRestatementFromRaw("cdcList_20210911",
"cdcList_20211203",
lstLabels=c("deaths_210911", "deaths_211203"),
labelBase=TRUE
) %>%
plotRestatementFromRaw(varNegTotal=c("state", "age"), makeProp=TRUE, makePropYears=2021)
# Create function for custom quarter-year
tmpCustomQuarter <- function(x, keyYear=2022)
ifelse(lubridate::year(x) %in% all_of(keyYear),
paste0(lubridate::year(x), "-Q", lubridate::quarter(x)),
lubridate::year(x)
)
calculateRestatementFromRaw("cdcList_20210911",
"cdcList_20211203",
lstLabels=c("deaths_210911", "deaths_211203")
) %>%
plotRestatementFromRaw(varNegTotal=c("state", "age"),
fnDateStack=function(x) tmpCustomQuarter(x, keyYear=2021)
)
Negative restatement of data was much less common, particularly among people under age 45, during a 3-month time period selected from 2021